Commit7a5c9ca

committed

Initialize shared memory copy of ckptXidEpoch correctly when not in recovery.

This bug was introduced by commit20d98ab,so backpatch this to 9.0-9.2 like that one.This fixes bug #6710, reported by Tarvi Pillessaar

1 parent4741e9a commit7a5c9caCopy full SHA for 7a5c9ca

File tree

-4

lines changed

-4

lines changed

Lines changed: 3 additions & 4 deletions

Original file line number	Diff line number	Diff line change
`@@ -6227,11 +6227,14 @@ StartupXLOG(void)`
`6227`	`6227`	`ereport(PANIC,`
`6228`	`6228`	`(errmsg("invalid next transaction ID")));`
`6229`	`6229`
	`6230`	`+/* initialize shared memory variables from the checkpoint record */`
`6230`	`6231`	`ShmemVariableCache->nextXid=checkPoint.nextXid;`
`6231`	`6232`	`ShmemVariableCache->nextOid=checkPoint.nextOid;`
`6232`	`6233`	`ShmemVariableCache->oidCount=0;`
`6233`	`6234`	`MultiXactSetNextMXact(checkPoint.nextMulti,checkPoint.nextMultiOffset);`
`6234`	`6235`	`SetTransactionIdLimit(checkPoint.oldestXid,checkPoint.oldestXidDB);`
	`6236`	`+XLogCtl->ckptXidEpoch=checkPoint.nextXidEpoch;`
	`6237`	`+XLogCtl->ckptXid=checkPoint.nextXid;`
`6235`	`6238`
`6236`	`6239`	`/*`
`6237`	`6240`	`* We must replay WAL entries using the same TimeLineID they were created`
`@@ -6330,10 +6333,6 @@ StartupXLOG(void)`
`6330`	`6333`	`/* No need to hold ControlFileLock yet, we aren't up far enough */`
`6331`	`6334`	`UpdateControlFile();`
`6332`	`6335`
`6333`		`-/* initialize shared-memory copy of latest checkpoint XID/epoch */`
`6334`		`-XLogCtl->ckptXidEpoch=ControlFile->checkPointCopy.nextXidEpoch;`
`6335`		`-XLogCtl->ckptXid=ControlFile->checkPointCopy.nextXid;`
`6336`		`-`
`6337`	`6336`	`/* initialize our local copy of minRecoveryPoint */`
`6338`	`6337`	`minRecoveryPoint=ControlFile->minRecoveryPoint;`
`6339`	`6338`

Comments

(0)